A Carnot-type engine with a changing phase working fluid is modelled and optimized. The engine is externally heated by a waste heat in order to valorize it. The waste heat is a flue gas at 350 degrees C, corresponding to a usual temperature of industrial furnace effluents. The objective function is the maximization of the net power output. In the numerical application, the working fluid is water. The optimization variables are the working fluid temperature of vaporization, its temperature of condensation and the allocation of a total thermal transfer area between the boiler and the condenser. This article aims to define an upper bound for waste heat to power conversion, depending on the waste heat temperature and its available mass flow rate, on the working fluid used and on the exchangers' size. The methodology used can be applied to optimize other externally heated engines. Different control volumes are selected for the analysis: the adiabatic converter, the non-adiabatic converter, the system (the converter and the exchangers) and the system placed in the environment. The impact of the choice of the control volume on optimization results, global efficiency and entropy production is studied. It appears that the optimal heating and cooling temperatures are the same in all sub-systems, but the optimal allocation of the total thermal transfer area differs slightly. However, in all cases the optimal condenser thermal transfer area is larger than the optimal boiler thermal transfer area. This imbalance corresponds to the need to evacuate the mechanical energy degradation into heat. It also appears that the control volume increase implies a decrease of the first law efficiency. The entropy analysis allows an identification and a quantification of the different contributions to entropy production. Then, a sensitivity analysis to the relevant parameters is carried out. The relevant parameters are the heating fluid temperature and mass flow rate, the total thermal transfer area, the compressor efficiency and the turbine efficiency.